This invention is in the area of biologically active nucleosides, and specifically includes a method for the resolution of nucleoside enantiomers, including 1,3-oxathiolane nucleosides, and antiviral compositions that include the enantiomerically enriched 1,3-oxathiolane nucleosides, (-) and (+)-2-hydroxymethyl-5-(5-fluorocytosin-1-yl)-1,3-oxathiolane ("FTC").
In 1981, acquired immune deficiency syndrome (AIDS) was identified as a disease that severely compromises the human immune system, and that almost without exception leads to death. In 1983, the etiological cause of AIDS was determined to be the human immunodeficiency virus (HIV). In December, 1990, the World Health Organization estimated that between 8 and 10 million people worldwide were infected with HIV, and of that number, between 1,000,000 and 1,400,000 were in the U.S.
In 1985, it was reported that the synthetic nucleoside 3'-azido-3'-deoxythymidine (AZT) inhibits the replication of human immunodeficiency virus type 1. Since then, a number of other synthetic nucleosides, including 2',3'-dideoxyinosine (DDI); 2',3'-dideoxycytidine (DDC), 3'-fluoro-3'-deoxythymidine (FLT), 2',3'-dideoxy-2',3'-didehydrothymidine (D4T), and 3'-azido-2',3'-dideoxyuridine (AZDU), have been proven to be effective against HIV. A number of other 2',3'-dideoxynucleosides have been demonstrated to inhibit the growth of a variety of other viruses in vitro. It appears that, after cellular phosphorylation to the 5'-triphosphate by cellular kinases, these synthetic nucleosides are incorporated into a growing strand of viral DNA, causing chain termination due to the absence of the 3'-hydroxyl group.
In its triphosphate form, 3'-azido-3'-deoxythymidine is a potent inhibitor of HIV reverse transcriptase and has been approved by the FDA for the treatment of AIDS. However, the benefits of AZT must be weighed against the severe adverse reactions of bone marrow suppression, nausea, myalgia, insomnia, severe headaches, anemia, peripheral neuropathy, and seizures. These adverse side effects often occur immediately after treatment begins, whereas a minimum of six weeks of therapy is necessary to realize AZT's benefits. DDI, which has recently been approved by an FDA Committee for the treatment of AIDS, is also associated with a number of side effects, including sporadic pancreatis and peripheral neuropathy.
Both DDC and D4T are potent inhibitors of HIV replication with activities comparable (D4T) or superior (DDC) to AZT. However, both DDC and D4T are not efficiently converted to the corresponding 5'-triphosphates in vivo. Both compounds are also toxic and can cause peripheral neuropathies in humans.
The success of various 2',3'-dideoxynucleosides in inhibiting the replication of HIV in vivo or in vitro has led a number of researchers to design and test nucleosides that substitute a heteroatom for the carbon atom at the 3'-position of the nucleoside. Norbeck, et al., disclose that (.+-.)-1-[(2.beta.,4.beta.)-2-(hydroxymethyl)-4-dioxolanyl]thymine (referred to below as (.+-.)-dioxolane-T) exhibits a modest activity against HIV (EC.sub.50 of 20 .mu.m in ATH8 cells), and is not toxic to uninfected control cells at a concentration of 200 .mu.m. Tetrahedron Letters 30 (46), 6246, (1989).
European Patent Application Publication No. 0 382 526 filed by IAF Biochem International, Inc. discloses a number of substituted 1,3-oxathiolanes with antiviral activity, and specifically reports that the racemic mixture (about the C4'-position) of the C1'-.beta. isomer of 2-hydroxymethyl-5-(cytosin-1-yl)-1,3-oxathiolane (referred to below as (.+-.)-BCH-189) has approximately the same activity against HIV as AZT, and no cellular toxicity at therapeutic levels. (.+-.)-BCH-189 has also been found to inhibit the replication of AZT-resistant, HIV isolates from patients who have been treated with AZT for longer than 36 weeks.
To market a nucleoside for pharmaceutical purposes, it must not only be efficacious with low toxicity, it must also be cost effective to manufacture. An extensive amount of research and development has been directed toward new, low cost processes for large scale nucleoside production. 2',3'-Dideoxynucleosides are currently prepared by either of two routes: derivatization of an intact nucleoside or condensation of a derivatized sugar moiety with a heterocyclic base. Although there are numerous disadvantages associated with obtaining new nucleoside analogues by modifying intact nucleosides, a major advantage of this approach is that the appropriate absolute stereochemistry has already been set by nature. However, this approach cannot be used in the production of nucleosides that contain either nonnaturally occurring bases or nonnaturally occurring carbohydrate moieties (and which therefore are not prepared from intact nucleosides), such as 1,3-oxathiolane nucleosides and 1,3-dioxolane nucleosides.
When condensing a carbohydrate or carbohydrate-like moiety with a heterocyclic base to form a synthetic nucleoside, a nucleoside is produced that has two chiral centers (at the C1' and C4'-positions), and thus exists as a diasteromeric pair. Each diastereomer exists as a set of enantiomers. Therefore, the product is a mixture of four enantiomers.
It is often found that nucleosides with nonnaturally-occurring stereochemistry in either the C1' or the C4'-positions are less active than the same nucleoside with the stereochemistry as set by nature. For example, Carter, et al., have reported that the concentration of the (-)-enantiomer of carbovir (2',3'-didehydro-2',3'-dideoxyguanosine) required to reduce the reverse transcriptase activity by 50% (EC.sub.50) is 0.8 .mu.M, whereas the EC.sub.50 for the (+)-enantiomer of carbovir is greater than 60 .mu.M. Antimicrobial Agents and Chemotherapy, 34:6, 1297-1300 (June 1990).
U.S. Ser. No. 07/659,760 discloses that 1,3-oxathiolane and 1,3-dioxolane nucleosides can be prepared with high diastereoselectivity (high percentage of nucleoside with a .beta. configuration of the bond from the C1'-carbon to the heterocyclic base) by careful selection of the Lewis acid used in the condensation process. It was discovered that condensation of a 1,3-oxathiolane nucleoside with a base occurs with almost complete .beta.-stereospecificity when stannic chloride is used as the condensation catalyst, and condensation of 1,3-dioxolane with a base occurs with almost complete .beta.-stereospecificity when various chlorotitanium catalysts are employed. Other Lewis acids provide low (or no) C1'-.beta. selectivity or simply fail to catalyze the reactions.
There remains a strong need to provide a cost effective, commercially viable method to obtain .beta.-stereospecificity of synthetic nucleosides prepared by condensing a carbohydrate-like moiety with a base. This is important because it is likely that many synthetic nucleoside inhibitors of viral replication now emerging from academic and commercial laboratories will require resolution. An economical and facile method for resolving racemic mixtures of nucleosides would greatly facilitate antiviral research and ultimately, commercial manufacture. Further, resolution of racemic mixtures of nucleosides may provide a route to increase the activity of synthetic nucleosides by eliminating or minimizing the undesired enantiomer.
Therefore, it is an object of the present invention to provide a method for the resolution of racemic mixtures of nucleosides.
It is another object of the present invention to provide enantiomerically enriched 1,3-oxathiolane nucleosides.
It is still another object of the present invention to provide enantiomerically enriched 1,3-oxathiolane nucleosides with significant antiviral activity and low toxicity.